Over six frightening months, a deadly germ untreatable by most antibiotics spread in the nation's leading research hospital. Pretty soon, a patient a week was catching the bug. Scientists at the National Institutes of Health locked down patients, cleaned with bleach, even ripped out plumbing - and still the germ persisted.
By the end, 18 people harbored the dangerous germ, and six died of bloodstream infections from it. Another five made it through the outbreak only to die from the diseases that brought them to NIH's world-famous campus in the first place.
It took gene detectives teasing apart the bacteria's DNA to solve the germ's wily spread, a CSI-like saga with lessons for hospitals everywhere as they struggle to contain the growing threat of superbugs.
It all stemmed from a single patient carrying a fairly new superbug known as KPC - Klebsiella pneumoniae that resists treatment by one of the last lines of defense, antibiotics called carbapenems.
"We never want this to happen again," said Dr. Tara Palmore, deputy hospital epidemiologist at the NIH Clinical Center.
Infections at health care facilities are one of the nation's leading causes of preventable death, claiming an estimated 99,000 lives a year. They're something of a silent killer, as hospitals fearful of lawsuits don't like to publicly reveal when they outfox infection control - yet no hospital is immune.
Wednesday, government researchers published an unusually candid account of last year's outbreak, with some advice: Fast sequencing of a germ's genome, its full DNA, may be essential. It can reveal how drug-resistant bacteria are spreading so that doctors can protect other patients.
"This is not an easy story to tell," said Dr. Julie Segre, a senior investigator at NIH's National Human Genome Research Institute. She led the genetic sleuthing that found the bug hiding in sink drains and, most chilling, even in a ventilator that had been cleaned with bleach.
Infection-control specialists at other hospitals called this detailed anatomy of an outbreak, published in the journal Science Translational Medicine, important to share.
"They were able to demonstrate that this sneaky little bug was able to stay alive and get transmitted in ways they hadn't quite predicted before they had the detailed genetic information," said Dr. Sara Cosgrove, associate hospital epidemiologist at Johns Hopkins University. "It's very revealing."
"Absolutely this could happen in any hospital," said Dr. Deverick Anderson, co-director of a Duke University infection control network that advises smaller community hospitals.
"This is really exciting stuff, cutting-edge technology, to try and better understand how these infections get spread," he added. That in turn may lead to new protections, important because "there's something that's very, very wrong about going to a hospital and becoming more ill."
Normally, the Klebsiella bacteria live in human intestines and don't harm people with healthy immune systems. But the multidrug-resistant strain named KPC has emerged over the past decade to become a fast-growing threat in intensive care units, spreading easily between very ill people and killing half of those it sickens. Worse, people can carry KPC without symptoms unless the germs slip into the urinary tract or bloodstream - theirs or the person's in the next bed - through a catheter or surgical wound.
The 243-bed NIH Clinical Center, in Washington's suburbs, is a unique hospital, only treating people enrolled in government research studies.
So on June 13, 2011, a research nurse carefully checked the medical records as a New York City hospital transferred a study participant who had become critically ill with a rare lung disease. The nurse found that the patient had KPC as well.
The woman went into strict isolation: Everyone entering her room donned a protective gown and gloves and rigorously washed their hands. Her medical equipment got special decontamination. All other patients in the ICU had their throats and groins tested regularly to see if the bug was spreading.
All seemed OK. The woman recovered, and went home on July 15.
Fast forward three weeks. Now a man with cancer has KPC despite never crossing paths with Patient No. 1. Ten days later, a woman with an immune disease fell ill, too. Both died of the infection.
Did they arrive carrying their own KPC bacteria, or did that first patient's germ somehow escape into the hospital? Standard tests couldn't tell. Segre, the geneticist, turned to DNA.
As bacteria multiply, mistakes appear and are repaired in their genetic code. Sequencing that genome allowed Segre to follow differences in single genetic letters like a trail of the germ's transmission and evolution.
Sure enough, the KPC originated from the New York patient despite NIH's precautions. Testing bacteria from the 17 additional patients who ultimately caught it shows the KPC was transmitted three separate times from Patient No. 1, and then spread more widely.
Even this sophisticated technology couldn't prove exactly how transmission occurred. But it turns out that Patient 3 had been in the ICU at the same time as the New York woman and really was the next infected, silently carrying the bug longer before becoming sick. That was enough time for Patient 3's infection to spread to Patient 2, who just got sick faster.
Meanwhile, NIH was making big changes. All the ICU patients underwent more invasive testing, using rectal swabs, to check for silent germ carriers. A new wall created a separate ICU to house them. Doctors, nurses, even janitors assigned there could work nowhere else, and monitors were paid to make sure everyone followed infection-control rules.
Yet a patient a week was either becoming infected or found to be a silent carrier of the same KPC strain.
"Honestly, we were very scared at that point," Segre recalled.
Test after test never found the bug on hospital workers' hands. Tainted objects like the ventilator couldn't be ruled out - but NIH adopted more complex and expensive decontamination, using robot-like machines to spray germ-killing hydrogen peroxide into the tiniest of crevices in all affected rooms and equipment.
Still, November brought more bad news: The outbreak strain had escaped the ICU, as two patients who'd never been there now were carrying it. A new isolation room was built, and all 200-plus patients in the hospital started undergoing rectal testing.
The outbreak now is over, the last carrier found in December. But NIH isn't dropping its guard. The isolation room remains, used every time one of the seven outbreak survivors returns to the hospital for their ongoing research studies - because they still carry the strain. Those rectal tests continue, hospital-wide once a month, to be sure no new KPC strain sneaks in.
Bacterial sequencing is becoming fast and cheap enough for most large hospitals to use during tough outbreaks, said Dr. Lance Peterson, microbiology and infectious disease director at NorthShore University HealthSystem in Evanston, Ill.
But another lesson is how much it takes to guard against these bugs sneaking in in the first place. Peterson said his hospital does weekly rectal testing of every ICU patient as a precaution.
"There's better technology becoming available for your hospital to prevent these bacteria from spreading, and this is what you should expect from your hospital," he said.